Study On Compensation Transformer Current Differential Protection

Ratio restrain and second harmonic escapement are the traditional differential protection. Differential protection malfunction are always caused by inrush current, over excitation and CT saturation. In fact, the phenomena are all belong to the core saturation of transformer. When exciting current is increasing, second harmonic escapement can play an important role in restraining inrush current; however, it is still not perfect. As electromagnetic transformer will gradually being replaced by electronic transformer, it is not necessary be influenced by its saturation phenomenon. To comply with power network development, this paper studies a new modified differential protection, which only considers the transformer saturation. It compensates exciting current to the differential current, and it needn't adopt any method to lock inrush current.When the transformer saturation has been detected, the exciting current is quite large. Suppose it is known for magnetization curve of transformer, substitute differential current as the exciting current into the magnetization curve. Then, compute flux linkage of a moment t, and use the flux linkage to get back to the magnetization curve, accordingly, the exciting current of a moment t can be compute. Finally, compensate the exciting current into the differential current, so that it can make reliable protection.The protection schemes need to solve the two key issues. They are the transformer winding parameter identification and singularity detection for transformer saturation. It can only get the sum of resistances or leakages for transformer windings by a short-circuit test, and the winding parameter can only be identified by Least-squares method during transient state. Ridge estimation is applied in this paper, for the first time. It can solve this morbid problem of the identification equation during the steady state, and the identification results have high precision. In contrast to all kinds of singularity detection methods, finally, a sliding-window iterative Discrete Fourier Transform is adopted. This method can meet the real-time for transformer saturation state detection with less computation and high anti-interference performances.The algorithm was validated by EMTP simulations and dynamic simulation experiments for its feasibility.